In response to anthropogenic emissions, hot days are projected to warm more than the mean day in the tropics. Climate models and theories have attributed enhanced hot day warming to reduced surface evaporative cooling during those days. However, we currently do not have a complete understanding of 1) the seasonal cycle of enhanced hot day warming, 2) the robustness of the signal across models, and 3) a mechanistic understanding of reduced evaporative cooling on hot days. Here we investigate how well proposed theories of the enhanced hot day warming explain the spatiotemporal variations in the magnitude of hot day warming across the CMIP6 ensemble. The results show warming of hot days are most pronounced during summertime. Additionally, models that project larger enhanced warming of hot days also project a larger decrease in surface evaporative cooling in those days. Both the spatiotemporal structure and the intermodel spread of enhanced hot day warming can be understood based on the climatological evapotranspiration regime of hot and mean days. Specifically, hot days warm more than the mean because hot days are in the water-limited regime (evaporation is sensitive to drying soil) whereas mean days are in the energy-limited regime (evaporation is insensitive to drying soil). We investigate an emergent constraint on enhanced warming of hot days by evaluating the spatiotemporal structure of evapotranspiration regimes simulated by models with observational datasets.